JuliaPy / JuliaPy/PyCall.jl

LLVM ERROR: Symbols not found: { ccall_PyByteArray_Type_105607, ccall_PyTuple_Type_105595, ccall_PyList_Type_105613 }

Offen
#924 2 Kommentare 0 Reaktionen 0 zugewiesene Personen Auf GitHub ansehen
Vorherrschende Sprache
Julia
Sterne
1.5k
Forks
186
PR-Merge-Kennzahlen
Keine gemergten PRs in 30 T.

Beschreibung

I'm trying to create a sysimage that includes `PyCall`.

Julia v1.5.4
PyCall v1.92.3

The build runs in an Ubuntu 18.04 docker container (it's part of the docker build process).

The sysimage creation goes on for about 96 seconds, and then fails with this error:

```
LLVM ERROR: Symbols not found: { ccall_PyByteArray_Type_105607, ccall_PyTuple_Type_105595, ccall_PyList_Type_105613 }
```

More output:
```
13:47:18 #74 63.50 [ Info: PackageCompiler: creating system image object file, this might take a while...
13:48:53 #74 159.0 LLVM ERROR: Symbols not found: { ccall_PyByteArray_Type_105607, ccall_PyTuple_Type_105595, ccall_PyList_Type_105613 }
13:48:53 #74 159.0
13:48:53 #74 159.2 ERROR: LoadError: failed process: Process(`/julia-1.5.4/bin/julia --color=yes --startup-file=no '--cpu-target=generic;sandybridge,-xsaveopt,clone_all;haswell,-rdrnd,base(1)' --sysimage=/julia-1.5.4/lib/julia/sys.so --project=/home/ubuntu/.julia/environments/v1.5 --output-o=/tmp/jl_ulcCLp.o -e '....
```

The strange thing is that it doesn't happen every time I run it. I've tried it maybe 6 times, and it fails about 50% of the time.

If it's any help, I recently upgraded from 1.92.2 to 1.92.3 but I didn't see anything in the diffs that could cause this.

Beitragsleitfaden

Für dieses Repository ist kein Beitragsleitfaden indexiert

Rechercherichtung

Start by reproducing sysimage creation with Julia v1.5.4, PyCall v1.92.3, and the Ubuntu 18.04 Docker environment, recording whether the LLVM symbol failure is intermittent. Compare successful and failed runs around PackageCompiler and PyCall initialization. Done means the reported symbols are resolved and sysimage creation completes reliably.

Vom Indexierungsmodell aus dem Issue-Text verfasst.

Bewertung

Tech-Stack
python
Bereich
build-system, devtools
Issue-Typ
Bug
Schwierigkeit
4/5
Geschätzter Aufwand
3-5 Tage
Aktivitätsstatus
Veraltet
Klarheit
Muss geklärt werden
Anfängerfreundlichkeit
25/100

Neue Issues direkt in Ihr Postfach

Eine kurze Übersicht über anfängerfreundliche GitHub-Issues.